My first Publication Agile-Data-Warehouse-Design-eBook | Page 44

How to Model a Data Warehouse
23
Figure 1-10
Order processing
ER Diagram
By looking at the ERD you can tell that customers may place orders for multiple
products at a time. The BEAM ✲ table records the same information, but the
example data also reveals the following:
Example data
models capture
more business
information than
Customers can be individuals, companies, and government bodies.
Products were sold yesterday.
Products have been sold for 10 years.
Products vary considerably in price.
Products can be bundles (made up of 2 products).
Customers can order the same product again on the same day.
Orders are processed in both dollars and pounds.
Orders can be for a single product or bulk quantities.
Discounts are recorded as percentages and money.
Additionally, by scanning the BEAM ✲ table you may have already guessed the type
of products that Pomegranate sells and come to some conclusions as to what sort
of company it is. Example data speaks volumes — wait until you hear what it says
about some of Pomegranate’s (fictional) staff!
ER models
Example data
speaks volumes!
Data Model Types
Agile dimensional modelers need to work with different types of models depend-
ing on the level of technical detail they are trying to capture or communicate and
the technical bias of their collaborators and target audience. Conceptual data
models (CDM) contain the least technical detail and are intended for exploring
data requirements with non-technical stakeholders. Logical data models (LDM)
allow modelers to record more technical details without going down to the data-
base specific level, while physical data models (PDM) are used by DBAs to create
database schemas for a specific DBMS. Table 1-2 shows the level of detail for each
model type, its target audience on a DW/BI project, and the BEAM ✲ diagram
types that support that level of modeling.
Conceptual, logical
and physical data
models provide
progressively more
technical detail for
more technical
audiences